GEO Under Lawyers' Professional Conduct Rules: What BRAO and BORA Actually Allow

Why GEO isn't new legal territory for law firms

The first time you hear that ChatGPT or Google AI Overviews 'recommends' clients to law firms, most lawyers ask the same question first: am I even allowed to let that happen? Yes. Lawyer advertising law has loosened considerably since the Federal Constitutional Court's 1987 Bastille decision — advertising is the rule now, prohibition the exception. Whatever applies to your website and your Google Business Profile applies, at its core, to AI answers as well.

Generative Engine Optimization just means writing your factual information so that language models can understand it, reproduce it accurately, and cite you as the source. You're not inventing a new kind of advertising — you're making information you're already allowed to publish machine-readable. The test is still Section 43b BRAO: advertising is permitted when it informs factually, in form and substance, about your professional activity, and isn't aimed at landing a specific individual mandate.

The framing matters here: the AI isn't advertising on your behalf — you're supplying the facts it draws on. That keeps you in the role of the lawyer who informs factually, and that role is exactly what professional conduct law protects.

Section 43b BRAO: the core rule and what it means for AI content

Section 43b BRAO sets three conditions worth memorizing: your information has to be factual in form and substance, it can't reference a specific mandate, and it can't be promotionally exaggerated. In practice, for GEO content: an FAQ answer like 'How does an unfair-dismissal claim proceed?' is fine, because it informs about your practice area. A line like 'Call now, your neighbor is suing you' crosses into case-specific solicitation and isn't allowed.

The tricky part with AI is that you don't fully control the output. A language model can sharpen or compress your factual statements when it repeats them. Under professional conduct law you're liable for what you publish yourself, not for every phrasing a model produces from it — so write your source text soberly and precisely enough that even a shortened version stays factual.

Skip superlatives like 'the best law firm,' 'guaranteed success,' or 'Germany's leading traffic-law attorney' unless you can prove them. Under Section 43b BRAO those are promotional claims, and language models tend to filter them out anyway, because they read as unsubstantiated.

Section 6 BORA: factual information is your safe harbor

Section 6 BORA says a lawyer may inform about their services and their person as long as the information is factual and professionally relevant. That's exactly where GEO and professional conduct law line up: factual, profession-related, verifiable information is what language models prefer to cite. Specialist-lawyer titles, practice areas, year of admission, languages, location, how a typical matter proceeds — all of that you can publish, and it's also what makes you quotable to an AI system.

Use professionally regulated terms correctly. Only call yourself a 'specialist lawyer for family law' (Fachanwalt) if you hold the title. 'Area of activity' (Tätigkeitsschwerpunkt) and 'area of interest' (Interessenschwerpunkt) are permitted under Section 7 BORA, but they're defined terms with a numeric cap. Use them precisely and you give the AI an unambiguous signal while staying compliant.

A common mistake: firms write something vague like 'We handle every area of law.' That's borderline misleading under professional conduct law and useless for GEO, because it gives the AI no clear competence signal to cite. Specific, honest statements about what you actually practice regularly perform better on both counts.

Client reviews and testimonials in AI answers

Language models and AI Overviews often pull from reviews on Google, Anwalt.de, or ProvenExpert. Reviews are permitted under professional conduct law as long as they're genuine and not steered in a misleading way. You can ask a satisfied client for a review. What you can't do is buy, fabricate, or incentivize reviews — that's unfair competition under the UWG and breaks the factuality requirement.

Watch success fees and outcome claims closely. If a client writes 'Mr. X won my case outright,' that's their opinion. It becomes a problem if you repeat that line in your own marketing copy as a guarantee of success. For GEO, genuine, specific reviews with real context read as more credible — and AI systems are more likely to treat them as a trustworthy signal.

There's a confidentiality angle too. Never name an identifiable client or case fact in a testimonial or case study without their express consent. Section 43a BRAO binds you to confidentiality, including in promotional text an AI later ingests.

The questions clients are already asking AI about your firm

People now ask language models questions they'd once have been hesitant to ask a lawyer directly: 'What does a divorce cost?', 'Am I owed severance?', 'How long does a fine proceeding take?' If your website answers those questions factually and precisely, you become the source the AI pulls from. That's entirely proper under professional conduct law, because you're informing generally about the legal situation and your practice, not soliciting an individual mandate.

Be careful with cost questions specifically. You can inform generally about the RVG (the statutory fee schedule), amounts in dispute, and cost risk. Flat fees are permitted where fee law allows them and they don't undercut the statutory minimum improperly. Keep the phrasing transparent — 'The statutory fee depends on the amount in dispute' is accurate, AI-friendly, and misleads no one.

Structure your content as clear question-and-answer pairs. Language models extract answers most easily from text that states a question and answers it in two or three concrete sentences — which happens to match the factuality requirement: short, correct, verifiable.

What professional conduct law actually prohibits in GEO

Once you know the boundaries, the no-go list is short: promotional touting, misleading statements, unfair comparisons naming other lawyers, and anything aimed at landing a specific mandate. If you build AI-optimized landing pages, keep unsubstantiated success claims, invented specializations, and disparaging remarks about other firms out of them.

One modern trap is AI-generated content on your own site. If you have ChatGPT draft text, have someone check it legally before you publish. Models sometimes invent statute sections, deadlines, or case citations. Publishing those errors can be misleading under both professional conduct law and competition law, and in the worst case creates a liability problem. Your duty of care doesn't stop at the prompt.

Collecting contact details through an AI chatbot on your site is regulated too. GDPR and confidentiality obligations apply from the first message. Make clear that the chatbot isn't legal advice and doesn't create an attorney-client relationship.

A practical roadmap for doing GEO safely under professional law

Start with an honest inventory of what you actually practice, and write a factual explainer for each area. Add structured data such as the LegalService schema from Schema.org, so search engines and AI systems can read your firm's details unambiguously. That's pure technical work, entirely uncontroversial under professional conduct law, and it noticeably improves how accurately models describe you.

Keep your Google Business Profile, your listings on lawyer-search directories, and your bar association record consistent. AI systems cross-check sources — a specialist title that shows up in one place and not another weakens your signal and, in extreme cases, can feed a misleading-advertising complaint. Consistency works as both a GEO lever and a compliance safeguard.

Keep a record of what you say where. If a colleague or your bar association questions a phrasing, you can show that your source content was factual. You're not responsible for what an AI generates on its own, but you are responsible for the source text you publish.

Conclusion: factual content is both the legal requirement and the best GEO strategy

The good news: what professional conduct law requires is nearly identical to what makes good GEO content. Factual, precise, verifiable, honest — that's the standard both Section 43b BRAO and modern language models judge your content against. Overstate your case and you lose twice: under professional conduct law, and in AI visibility, because models tend to filter out unsubstantiated claims.

Don't treat GEO as a legal gray area — treat it as the natural extension of advertising you're already allowed to do into the AI era. Your job is to describe your competence clearly and accurately enough that ChatGPT, Perplexity and Google AI Overviews treat you as a reliable source. Stay the factual lawyer you already are, and visibility follows as a consequence, not a risk.

If you're unsure about a specific format — a chatbot, an AI-assisted intake flow — ask your bar association before you launch it. Professional conduct law keeps evolving, but the underlying principle holds: factual information, yes; loud self-promotion, no.

Naming your focus areas in AI answers without crossing into specialist-title territory

When an AI answer names your firm, it often attaches a practice area — 'specializes in tenancy law' or 'handles traffic accidents.' Be careful here. Professional conduct law lets you name areas of interest and activity, but it requires a clear line between that and an actual specialist-lawyer designation. Only someone formally certified should have 'specialist lawyer for' (Fachanwalt für) attached to their name anywhere in their content. Everything else needs to read as a focus area, not a title.

In practice: write it consistently as 'focus area: family law' (Tätigkeitsschwerpunkt Familienrecht) rather than phrasing that implies a certification you don't hold. AI systems frequently repeat your wording verbatim, so if your source pages keep the distinction clean, the AI's output stays clean too. Check your legal notice, profile pages, and structured data for any uncertified specialist-lawyer title that's crept in — one sloppy mention can propagate through dozens of AI answers and draw a cease-and-desist letter from a colleague.

Success rates and win percentages: why "95 percent won" is a risk

Clients like asking AI systems for the 'best' or 'most successful' firm, which tempts some lawyers to pad their content with win rates so the model names them. But a blanket figure like '95 percent of cases won' is risky under professional conduct law — it's rarely provable, depends heavily on which cases you count, and can easily read as misleading advertising. You also have no control over how an AI system repeats a number like that.

Use verifiable, factual detail instead. A concrete, anonymized case example with the actual result achieved is permitted and still gives the AI usable context. Write 'In an unfair-dismissal matter, we secured a severance equal to two months' gross salary' rather than a bare percentage. That keeps you inside the factuality requirement while giving AI systems the specific, checkable material they need to produce a differentiated answer. Factual detail also just performs better than a superlative.

Who's liable when an AI gets your firm's facts wrong

A common worry: what if an AI system states something wrong about your firm — a practice area you don't handle, a title you don't hold? As a rule, you're not liable for an external system's independent errors. It becomes a real problem only if the false statement traces back to sloppy content you published, or if you know about the error and don't correct it.

That's why a regular self-check belongs in your GEO routine. Once a quarter, ask the major AI systems directly about your firm and read the answers critically. If you find a statement that's a problem under professional conduct law, log it with a date and screenshot, then fix the source content it came from. For errors that persist, use the provider's feedback mechanism. That record protects you twice — as evidence of diligence under professional conduct law, and as a working log for improving your online presence over time.